Characterization and modeling of creep in RF-MEMs tunable components and circuits

Hao-Han Hsu, Purdue University

Abstract

Creep in thin film metals is a potentially significant failure mechanism in RF-MEMS. Materials that creep exhibit a time-dependent response to a constant force and are widely employed in RF-MEMS, including thin film metals such as Au, Al, and Ni. Creep in RF-MEMS devices in this work is characterized through a highly-accurate capacitance-sensing setup, under a special bi-state bias condition. In particular, the devices are kept continuously biased (on-state) for up to 1,400 hours, but the bias voltage is momentarily removed for one minute every hour to record the off-state capacitance. The capacitance measurement uncertainty is less than 200 aF and the long-term stability is better than 4 fF. Furthermore, creep behavior under the same bi-state bias condition is investigated with a confocal microscope-based setup. A physics-based creep model for nanocrystalline nickel devices reveals that the observed creep deformation is dominated by Coble creep. Furthermore, a compact computer-aided-design (CAD) model that may be utilized to simulate the creep behavior of RF-MEMS varactors in RF circuits and sub-systems is developed based on the aforementioned measurements. This model is capable of calculating the long-term response of RF-MEMS devices to an arbitrary input waveform. It is experimentally-validated with measurements of Ni varactors that extend up to 760 hours. Its effectiveness is demonstrated with a tunable RF-MEMS resonator and an RF-MEMS phase shifter. The proposed creep model along with the measurements up to 1,400 hours can improve the understanding of long-term operation of RF-MEMS devices and may lead to more reliable designs. They may also play a key role in lifetime evaluation and prediction.

Degree

Ph.D.

Advisors

Peroulis, Purdue University.

Subject Area

Electrical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS